Experiment with activity Temperatures

[Kovox91]

According to this source the incubation time of dengue is roughly a week (simplified!).

Since the vector is a mosquto which usually have some critical temperature limits:

• critical low: no activity
• optimal: high activity
• critical high: lower activity
  (Yai study biology thei said!)

According to this review:

The lower temperature limit for Ae. aegypti is around 10 ◦C, a temperature below which mosquitoes become torpid and unable to move

which gives us the critical low.
Further

Ae. aegypti females were able to sustainably fly between 15◦C and 32◦C

e.g. optimal conditions. And lastly

flight was possible but only for short periods of time at extreme temperatures such as 10 ◦C and 35 ◦C.

Further, the active times of Ae. aegypti are in the morning and in the early evening (can't find the citation right now.

This leads me to believe that we can both simplify the temperature features AND give them more meaning. I suggest the final result would be a categorical feature with three states: critical low, optimal, critical high (this could be further refined but i think its a good starting point). ANd of course it would be lagged.
I also though about two lagged temperature features, one for infecting and one for reproduction?

I'm not sure if we should use the average or min and max temp, initially I though about min and max but since the average temperature is reached around 10 in the morning (roughly at least in our latitude) i guess that the average might be better?

I'll see if i can find the time to properly read the review, especially for the reproduction and infection split (might also be the same) so far I only skimmed the introduction but that should at least give us a good starting point, we can fine-tune the exact numbers later.

[Kovox91]

OKay, apparently the diurnal temperature range (DTR) is also important for their behaviour and we also have that feature...
Also there are papers on modelling temperature effects on dengue fever here but that really feels like cheating?

[Kovox91]

Yang et al. [64] used mathematical modeling (i.e., population dynamics theory) to determine that 29.2 ◦C was the optimal temperature to produce the highest amount of offspring in Ae. aegypti mosquitoes and maintain a viable mosquito population.

Man, there is so much in this review...

[Kovox91]

ok, one last thing before i go to bed.
In this paper they used models to ascertain the optimal temperatures. and came to the following conclusions:

the optimal range of temperature for survival of the adult mosquitoes is 15<T<30°C, while for the aquatic stages, we have 15<T<35°C. However, the temperature at which quick transitions occur in the aquatic stages is around 26°C. Therefore, the effect of temperature is distinct in different stages of the mosquito’s life-cycle.

And it goes on afterwards.
I guess we'll need to learn about the timings of ae.aegypti development and then make different lagged temperature optima (early development, later development, infection/mating)?

[Kovox91]

ok, I spent some time thinking about it and i'm almost sure that we dont need to do the whole switch to categorical features because the tree will figure the thresholds out on its own
so we rather need to focus on development timings and lag respectively

[Kovox91]

I will re-open this as I think there might be some merit to do more of this later on, for now I guess there are bigger problems